Influence of gaps on roof insulation performance

by nithya_caleb | September 20, 2018 12:00 am

Images courtesy Soprema[1]
Images courtesy Soprema

By Jean-François Côté, PhD, OCQ

As the construction industry evolves, product development and implementation of energy efficient solutions increase building performance. Insulation materials and assemblies, including roof insulation, are part of this ongoing process. Increased understanding of the dynamics of low-slope roof assemblies has helped design better performing roofs, from a thermal point of view, as well as other considerations such as moisture management and resistance to wind uplift.

The recommended practice in the roofing industry is to use two layers of insulation and stagger the joints. By doing this, if gaps open between insulation boards, the joints in the first layer will not be positioned directly over joints in the second layer. Consequently, the thermal efficiency of the roof assembly is considered to be less penalized when gaps appear because an insulation layer is always covering a gap. However, this does not prevent insulation from shrinking, so there could still be some appearance of gaps and different gap widths in those insulation materials.

What is the impact of these insulation gaps in the thermal efficiency of roof assemblies? Experiments were performed by RDH Building Science, National Research Council Canada (NRC), and a manufacturer of insulation materials in their respective laboratories to better answer the question. To define parameters for the testing, parameters by which insulation materials shrink needed to be first considered.

Thermal expansion and contraction

Figure 1: Thermal contraction from 23 C (73 F) to –15 C (5 F) for different types of insulation materials.[2]
Figure 1: Thermal contraction from 23 C (73 F) to –15 C (5 F) for different types of insulation materials.

All materials have a coefficient of thermal expansion. This behaviour is inherent in the nature of materials. When a material is heated, it increases in dimension (expansion) and decreases (contraction) when cooled. The thermal expansion and contraction is a totally reversible phenomenon. The dimensions of a material at a given temperature remain unchanged even after cycles of exposure to colder or warmer temperatures. Some materials exhibit larger change in dimensions than others, and the coefficient of thermal expansion is a number used to quantify this change. Further, the change in dimension may not be the same in all directions for a given material.

Specimens of three types of insulation materials (polyisocyanurate [polyiso], expanded polystyrene [EPS], and stone wool [SW]) were cut to a size of 600 x 600 mm (24 x 24 in.) at room temperature. They were then cooled to –15 C
(5 F) and specimen dimensions were measured at –15 C. Figure 1 shows the average percentage of shrinkage in width and length. Unsurprisingly, SW shows the lowest shrinkage of the group by at least an order of magnitude. Both plastic foam insulation materials (polyiso and EPS) show similar results with polyiso being the worst case in width at 0.158 per cent shrinkage. If this is transposed to a roof employing 1.2 x 1.2-m (4 x 4-ft) insulation boards, thermal expansion and contraction can create relatively small gaps of 1.9 mm (75 mils) in size between polyiso boards.

Dimensional stability

Staggered polyiso system (top) and shiplap polyiso board (bottom) differ in the mechanism for heat loss even if total thickness is identical. Images courtesy National Research Council Canada[3]
Staggered polyiso system (left) and shiplap polyiso board (right) differ in the mechanism for heat loss even if total thickness is identical.
Images courtesy National Research Council Canada

[4] Dimensional stability is sometimes confused with thermal expansion. Dimensional stability is a property found in product standards for insulation materials as well as for other roofing or building enclosure materials. It is a test for measuring the permanent deformation of a material after exposure to extreme conditions. In the case of insulation materials, dimensional stability is technology dependent, and has to be tested according to the product specification covering the type of insulation. In all cases, dimensions of the specimen are measured before and after the exposure period, once it has reached room temperature again. For EPS, the Canadian standard is Underwriters Laboratories of Canada (CAN/ULC) S701.1, Standard for Thermal Insulation, Polystyrene Boards. This standard requires EPS to be tested for dimensional stability by exposing a specimen for seven days at 70 C (158 F) (dark oven, no specific moisture level), and then letting it cool down to room temperature and measuring any change in dimension. The standard limits the change in dimension for EPS to 1.5 per cent.

In the case of polyiso, the standard of reference in Canada, CAN/ULC S704.1, Standard for Thermal Insulation, Polyurethane And Polyisocyanurate, Boards, Faced, has different requirements. Polyiso must be exposed to three different conditions when testing for dimensional stability. Those three conditions are independent and require a different specimen for the test. The maximum allowed change in dimension is 2 per cent. The first condition is a cold temperature exposure (seven days at –29 C [–20 F]); the second is a longer, dry heat exposure (28 days at 80 C [176 F]); and the third is a long, humid heat exposure (28 days at 70 C [158 F] and 97 per cent relative humidity [RH]).

Dimensional stability tests need to be performed by the manufacturers to document compliance to the product standard. Although the standard allows up to 2 per cent change in dimensions, typically less than 0.5 per cent change is measured for good quality polyiso. However, if design professionals take the maximum 2 per cent change, the potential net effect is 24 mm (0.95 in.) of shrinkage for a 1.2 x 1.2-m board.

Shrinkage due to moisture

Mechanically attached roof assemblies may be subjected to moisture accumulation within the insulation layer under the roof membrane. When the membrane billows due to uplift pressure generated by wind, it acts as a diaphragm and pulls air from the interior, conditioned space to the unconditioned space in the insulation layer. When the deck-level vapour barrier is not perfectly airtight, this air brings moisture that condenses at colder temperatures on the underface of the roof membrane. Even if the moisture is allowed to escape the roof assembly during the summer months, Molleti et al. (2015) showed insulation materials had been slightly affected and small gaps (up to 13 mm [½ in.] in the case of polyiso) appeared between insulation boards (For more information, refer to “Impact of Air Intrusion on Moisture Performance of Seam-Fastened Mechanically Attached Roofing Systems” by S. Molleti, B. Baskaran, P. Kalinger, M. Graham, J.-F. Côté, J. Malpezzi, and J. Schwetz, presented in 2017 at the 15th Canadian Conference on Building Science and Technology).

Heat flow meter testing

To evaluate the impact of the presence of gaps between insulation boards to the thermal performance of roof assemblies, two types of testing—small-scale, heat flow meter testing as well as a larger, hot box testing—were performed.

The heat flow meter is a small-scale test routinely used to measure R-value of materials. This test is referenced in all insulation standards for the determination of their R-value. Testing was done on polyiso according to ASTM C518, Standard Test Method for Steady-State Thermal Transmission Properties by Means of the Heat Flow Meter Apparatus. The heat flow meter has an insulation specimen positioned between cold and hot plates. Since there is a difference in temperature of the plates, a heat flow is induced within the insulation material. This heat flow is measured and converted into thermal resistance (R-value) of the specimen. The temperature differential is essential for the test because without it, there would be no heat flow through the specimen, hence it would be impossible to report a R-value. Consequently, the R-value obtained from heat flow meter testing is expressed with respect to a mean temperature, which is the average temperature between the cold and hot plates.

Specimens of polyiso with thickness ranging from 63 to 150 mm (2.5 to 6 in.) were tested in this study. Specimens measuring 300 x 300 mm (12 x 12 in.) were used in the heat flow meter except for 150-mm thick specimens, which required larger specimens (600 x 600 mm). Tests were performed at 24 C (75 F) mean temperature (cold plate at 13 C [55 F] and hot plate at 35 C [95 F]).

Figure 2: Polyisocyanurate (polyiso) specimens with straight (top) and shiplap (bottom) joints with 15-mm (591-mils) lip size.[5]
Figure 2: Polyisocyanurate (polyiso) specimens with straight (top) and shiplap (bottom) joints with 15-mm (591-mils) lip size.

A reference test was done on an integral specimen (a full specimen without a joint), to which all other tests would be compared. Testing continued using specimens with joints and creating gaps in the specimens. Straight and shiplap joints (with 15 mm [591 mils] lip size) were compared (Figure 2). Closed joints as well as joints with gaps of up to 25 mm (1 in.) were tested for comparative purposes up to the maximum shrinkage allowed by the standards, even though such large gaps are rarely seen in the field.

Specimens measured with a joint needed their R-value to be normalized. In the heat flow meter equipment, the joint is present on a 300-mm wide piece of insulation. On a roof, the same joint would be present for a 1220-mm (48-in.) wide board, so the joint’s influence on the overall thermal performance is greater for the heat flow meter result than it is on the roof. The R-values obtained from the heat flow meter were therefore normalized for a 1.2 x 1.2-m (4 x 4-ft) board size using an equation found in Burch et al. (1987) and assuming only a one-dimensional heat flow (ignoring any convection, conduction, and side effects). For more information, read the 1987 paper “A Heat Transfer Analysis of Metal Fasteners in Low-Slope Roofs” by D. Burch, P. Shoback, and K. Cavanaugh for selected technical papers (STP) 959 Roofing Research and Standards Development, edited by R.A. Critchell for ASTM International.

Normalized results of the heat flow meter testing are presented in Figure 3 . Reference R-values for the integral specimen increase from 16.4 to 40.7 depending on thickness, predictably. Also indicated are the R-values of specimens with joints (but closed gap), and with gaps up to 25 mm (1 in.), for both straight and shiplap joints.

Figure 3: Heat flow meter test R-value results.[6]
Figure 3: Heat flow meter test R-value results.

As expected, the R-values obtained on specimens with closed joints are lower than the reference. The difference is not large, reaching no more than 10 per cent reduction over the reference value. Interestingly, results seem to indicate straight joints perform better than shiplap joints when no gap is present. However, the largest difference is seen with the 63-mm (2.5-in.) thickness and is below 2 per cent, which cannot be considered meaningful. When gaps open to 6 mm (236 mils), results begin to favour the shiplap joint configuration as two results out of three indicate a better retention of R-value for shiplap over straight joints. This is especially true at the thickness of 152 mm (6 in.) where the shiplap specimen retained 83.4 per cent of the reference R-value and the straight joint specimen retained only 65.3 per cent. At 13 mm (½ in.) gap size, the dominance of shiplap configuration is obvious (89 over 75 per cent retention for straight joint). If gaps open more than 13 mm, the shiplap no longer functions as such (current specimens had a 15-mm [591-mils] lip) and both joints provide about the same R-value despite a little advantage for the shiplap (63 per cent retention compared to 61 per cent for straight joint).

These results are only indicative because the straight joint configuration is not representative of reality. As mentioned earlier, there are usually two layers of insulation on top of one another in a roof system. The joints are staggered so if a gap opens between boards, the gap on the top board is never positioned over the gap on the bottom board.

Figure 4: A calibrated hot box at National Research Council Canada’s (NRC’s) office in Ottawa. The top chamber can be lowered onto the table where the specimen is constructed. Photo courtesy National Research Council Canada[7]
Figure 4: A calibrated hot box at National Research Council Canada’s (NRC’s) office in Ottawa. The top chamber can be lowered onto the table where the specimen is constructed.
Photo courtesy National Research Council Canada

Hot box testing

The next part of the study was performed with full boards by NRC in Ottawa with a calibrated hot box. The equipment measures 2.4 x 6.1 m (8 x 20 ft) and replicates conditions acting on the roofing assembly at an intermediate scale. It is composed of two chambers and the test specimen is exposed to a temperature difference between the top and bottom chambers. The energy required to maintain both chambers at constant temperature conditions is measured and converted into an effective R-value. The testing was conducted according to the ASTM C1363, Standard Test Method for Thermal Performance of Building Materials and Envelope Assemblies by Means of a Hot Box Apparatus, specification for calibrated hot box (Figure 4).

For this study, the top chamber temperature was kept at 4.5 C (40 F) and the bottom chamber temperature was maintained at 41.5 C (107 F) (23 C [73 F] mean temperature and 37 C [99 F] temperature difference). A vapour impermeable air barrier was installed under the insulation boards to avoid any airflow or moisture movement into the specimens.

Assembly 1 was composed of two layers of 63.5-mm (2.5-in.) polyiso boards with staggered joints (this can be considered the reference). Joints were spaced by 0.60-m (2ft) between the layers in both directions. Assembly 2 used the same configuration but introduced 13-mm (½-in.) gaps at every joint in both layers of insulation. Assembly 3 was constructed with a single layer of 127-mm (5-in.) polyiso with no-gap shiplap joints. Finally, Assembly 4 used a single layer of 127-mm polyiso with shiplap joints with an added 13-mm gap at all joints in both directions (the shiplap edge is present on all edges of the boards).

Figure 5: Hot box testing R-value results. Image courtesy Soprema[8]
Figure 5: Hot box testing R-value results. Image courtesy Soprema

The results of the hot box testing are presented in Figure 5. When gaps are closed, shiplap joints (Assembly 3) provide marginally better R-values (3 per cent) than staggered system (Assembly 1), which is opposite to what one would have seen in the small-scale testing. When the gap size is increased to 13 mm, the difference in R-value increases in favour of the shiplap configuration which retains 90 per cent of the reference R-value as opposed to a 85 per cent retention for staggered joints. This seems to indicate using shiplap polyiso in a roofing assembly may bring benefits in terms of thermal performance when gaps do not exceed the dimension of the shiplap lip. The impact of increasing the gap size further will be part of future studies.

The increase in performance observed for shiplap polyiso is consistent with claims from polyiso manufacturers that R-value per inch of polyiso increases with thickness. A single layer of polyiso is composed of insulation foam between two facer materials. At an equivalent thickness, two layers of polyiso introduce two additional facer materials, which are not as thermally efficient as the thickness of foam they replace. Field trials confirmed an increase in speed of installation for shiplap joints over staggered systems (not so much for mechanically attached assemblies). Changing practices for roof insulation installation will not happen tomorrow. Polyiso has generally not been used in large thicknesses in Canada because failures have been observed where thick boards showed excessive cupping, curling, or unacceptable variations in thickness. Polyiso manufacturers capable of providing stable thick boards with shiplap edges might provide tangible benefits to the industry.

This study confirmed the presence of gaps between boards of insulation will cause a reduction in R-value. The different parameters that could cause insulation boards to shrink (coefficient of thermal expansion, dimensional stability, and moisture) may create gaps up to 25 mm, although typical dimensional stability data shows 13 mm is a more realistic maximum gap size for polyiso. Under these conditions, gaps created between polyiso boards with shiplap edges cause a smaller reduction (10 per cent) of the thermal performance of the assembly compared to dual layers with staggered joints (15 per cent).

[9]Jean-François Côté, PhD, OCQ, is the director of standards and scientific affairs for Soprema. A registered chemist in Québec, he chairs the Canadian Standards Association (CSA) A123.23 technical committee on bituminous roofing materials. Côté is also an active member of ASTM International, Asphalt Roofing Manufacturers Association (ARMA), Polyisocyanurate Insulation Manufacturers Association (PIMA), and Underwriters Laboratories of Canada (ULC). He can be reached at jfcote@soprema.ca[10].

Endnotes:
  1. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/IMG_5558-.jpg
  2. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/Figure-1-1.jpg
  3. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/Shiplap.jpg
  4. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/ISO.jpg
  5. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/roof_Figure-2.doc.jpg
  6. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/Figure-3.jpg
  7. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/Hot.jpg
  8. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/Figure-5.jpg
  9. [Image]: https://www.constructioncanada.net/wp-content/uploads/2018/09/JFCOTE_Photo.jpg
  10. jfcote@soprema.ca: mailto:jfcote@soprema.ca

Source URL: https://www.constructioncanada.net/influence-of-gaps-on-roof-insulation-performance/